The present invention relates to a disk array apparatus which is configured to efficiently acquire a backup, and more particularly to a backup acquisition method and a disk array apparatus which can back up a disk area while sequentially reading the area without being conscious of a file system.
Generally, data recorded in a storage device must be periodically backed up using a backup device such as a magnetic tape device on the assumption that the data could be lost due to possible contingencies. Since an update to data being backed up, if made during a backup operation, can result in discrepancies in data, the data must be protected from being updated during the backup operation. However, in a system which is required to provide a high availability, it is difficult in view of the operation to stop updating data during a backup operation. Thus, for acquiring a backup, a snapshot of data, described below, is utilized to prevent the data from being updated, before the data is backed up.
Generally, a snapshot refers to an image of stored data at a certain time. Data subjected to the snapshot is called “original data.” Immediately after a snapshot is captured, the snapshot and original data physically share a storage area. For updating data after the capture of the snapshot, the data before the update is recorded in a different area (pre-update data storage area) before the original data in the storage area is updated. Also, for referencing the original data, data should be always read from the storage area of the original data. On the other hand, for referencing a snapshot, data which is not updated after the capture of the snapshot is read from the storage area of the original data, and data which has been updated after the capture of the snapshot is read from a different area in which data before the update is recorded.
As described above, the snapshot and original data can be accessed as logically independent different data, so that the snapshot is not updated even when the original data is updated.
In a snapshot which is captured for making a backup, an update to data after the capture of the snapshot involves preserving the data before the update in a different area (pre-update data storage area), and then updating the original data storage area. As a result, the storage area of the snapshot will mixedly exist in the original data storage area and the pre-update data storage area. For referencing the snapshot, required portions must be referenced astride both the areas, disadvantageously resulting in the inability to sequentially reading a single disk area for acquiring a backup.
Also, during the acquisition of a backup, frequent seek operations are required for searching the original data storage area and pre-update data storage area for a differential backup from a whole backup, disadvantageously causing an increase in the processing time and time-consuming operations.